dsh-revbayes-forge-pack-model-selection

Model Selection pack

Development fixture package dsh-revbayes-forge-pack-model-selection of the RevBayes Forge plugin (dsh-revbayes-forge).

English

• Pack id: model-selection
• Tool: revbayes_model_selection
• Maturity: golden

What it generates. Builds marginal-likelihood scripts that wrap a base script or base spec in a power posterior and attach a stepping-stone (steppingStoneSampler) or path-sampling (pathSampler) estimator. Steps (cats >= 2), burn-in, and sampling frequency are validated; the resulting runs are computationally expensive.

Analyses.

Analysis	Title	Canonical golden
stepping_stone	Stepping-stone	stepping_stone.Rev
path_sampling	Path sampling	path_sampling.Rev
marginal_likelihood	Marginal likelihood	—

Full spec fields, data requirements, and limitations: docs/en/model-packs.md.

Fixtures. Files under fixtures/ are synthetic small datasets shipped with the repository; they back the golden scripts, canonical specs, and runtime checks used by npm test and npm run test:runtime.

Enablement. Enable this pack in the settings page (RevBayes Forge → Packs); no separate installation is needed.

---

中文

• Pack id：model-selection
• 工具：revbayes_model_selection
• 成熟度：golden

生成内容。 生成边际似然脚本：把 base 脚本或 base spec 包进 power posterior，并附加 stepping-stone（steppingStoneSampler）或 path-sampling（pathSampler）估计器。校验步数（cats >= 2）、burn-in 与采样频率；生成的计算开销较大。

分析列表。

Analysis	Title	Canonical golden
stepping_stone	Stepping-stone	stepping_stone.Rev
path_sampling	Path sampling	path_sampling.Rev
marginal_likelihood	Marginal likelihood	—

完整 spec 字段、数据要求与限制：docs/zh/model-packs.md。

夹具数据。 fixtures/ 下的文件均为随仓库分发的合成小数据，支撑 golden 脚本、canonical spec 与运行时检查，供 npm test 与 npm run test:runtime 使用。

启用方式。 在设置页（RevBayes Forge → Packs）勾选启用即可，无需单独安装。